Things to consider when choosing infrastructure.

With all the choice in the compute/storage market at the moment, choosing new infrastructure for your next project is not an easy task.

In my experience most customers (and many architects) think about the infrastructure coming up for replacement and look to do a “like for like” replacement with newer/faster technology.

An example of this would be a customer with a FC SAN running Oracle workloads where the customer or architect replaces the end of life Hybrid FC SAN with an All Flash FC SAN and continues running Oracle “as-is”.

Now I’m not saying there is anything wrong with that, however if we consider more than just the one workload, we may be able to achieve our business requirements with a more standardized and cost effective approach than having dedicated infrastructure for specific workloads.

So in this post, I am inviting you to consider the bigger picture.

If we take an example customer has the following workload requirements:

  1. Virtual Desktop (VDI)
  2. Virtualized Business Critical Applications (e.g.: SQL / Exchange)
  3. Long Term Archive (High Capacity, low IOPS)
  4. Business Continuity and Disaster Recovery

It is unlikely any one solution from any vendor is going to be the “best” in all areas as every solution has its pros and cons.

Regarding VDI, I would say most people would agree Hyperconverged Infrastructure (HCI) / Scale out type architectures are strong for VDI, however VDI can be successfully deployed on a traditional SAN/NAS solutions or using non shared local storage in the case of non-persistent desktops.

For vBCA, some people believe physical servers with JBOD storage is best for workloads like Exchange, and Physical + local SSD are best for Databases while many people are realising the benefits of virtualization of vBCA with shared storage such as SAN/NAS or on HCI.

For long term archive, cost per GB is generally one of if not the most critical factor where lots of trays of SATA storage connected to a small dual controller setup may be the most cost effective, whereas an All Flash array would be less likely considered in this use case.

For BC/DR, features such as a Storage Replication Adapter (SRA) for VMware Site Recovery Manager, a stretched cluster capability and some form of snapshot capability and replication would be typical requirements. Some newer technology can do per VM snapshots, whereas older style SAN/NAS technology may be per LUN, so newer technology would have an advantage here, but again, this doesn’t mean one tech should not be considered.

So what product do we choose for each workload type? The best of breed right?

Well, maybe not. Lets have a look at why you might not want to do that.

The below graph shows an example of 3 vendors being compared across the 4 categories I mentioned above being VDI, vBCA, Long Term Archive and BC/DR.

ExmapleGraph

The customer has determined that a score of 3 is required to meet their requirements so a solution failing to achieve a 3 or higher will not be considered (at least for that workload).

As we can see, for VDI Vendor B is the strongest, Vendor A second and Vendor C third, but when we compare BC/DR Vendor C is strongest followed by Vendor A and lastly Vendor B.

We can see for Long Term Archive Vendor A is the strongest with Vendor B and C tied for second place and finally for vBCA Vendor B is the strongest, Vendor A second and Vendor C third.

So if we chose the best vendor for each workload type (or the “Best of breed” solution) we would end up with three different vendors equipment.

  • VDI: Vendor B
  • Long Term Archive: Vendor A
  • BC/DR: Vendor C
  • vBCA: Vendor B

Is this a problem? Not necessarily but I would suggest that there are several things to consider including:

1. Having 3 different platforms to design/install/maintain

This means 3 different sets of requirements, constraints, risks, implications need to be considered.

Some large organisations may not consider this a problem, because they have a team for each area, but isn’t the fact the customer has to have multiple teams to manage infrastructure a problem in itself? Sounds like a significant (and potentially unnecessary) OPEX to me.

2. The best BC/DR solution does not meet the minimum requirements for the vBCA workloads.

In this example, the best BC/DR solution (Vendor C) is also the lowest rated for vBCA. As a result, Vendor C is not suitable for vBCA which means it should not be considered for BC/DR of vBCA. If Vendor C was used for BC/DR of the other workloads, then another product would need to be used for vBCA adding further cost/complexity to the environment.

3. Vendor A is the strongest at Long Term Archive, but has no interoperability with Vendor B and C

Due to the lack of interoperability, while Vendor A has the strongest Archiving solution, it is not suitable for this environment. In this example, the difference between the strongest Long Term Archive solution and the weakest is very small so Vendor B and C also meet the customers requirements.

 4. Multiple Silos of infrastructure may lead to inefficient use.

Just like in the days before Virtualization, we had the bulk of our servers CPU/RAM running at low utilization levels, we had our storage capacity carved up where we had lots of free space in one RAID pack but very little free space in others and we spent lots of time migrating workloads from LUN 1 to LUN 2 to free up capacity for new or existing workloads.

If we have 3 solutions, we may have many TB of available capacity in the VDI environment but be unable to share it with the Long Term Archiving. Or we may have lots of spare compute in VDI and be unable to share it with vBCA.

Now getting back to the graph, the below is the raw data.

rawdata

What we can see is:

  • Vendor B has the highest total (17.1)
  • Vendor A has the second highest total (14.8)
  • Vendor C has the lowest total (12)
  • Vendor C failed to meet the minimum requirements for VDI & vBCA
  • Vendor A and B met the minimum requirements for all areas

Let’s consider the impact of choosing Vendor B for all 4 workload types.

VDI – It was the highest rated, met the minimum requirements for the customer and is best of breed, so in this case Vendor B would be a solid choice.

vBCA – Again Vendor B was the highest rated, met the minimum requirements for the customer and is best of breed, so Vendor B would be a solid choice.

Long Term Archiving: Vendor B was equal last, but importantly met the customer requirements. Vendor A’s solution may have more features and higher performance, but as Vendor B met the requirements, the additional features and/or performance of Vendor A are not required. The difference between Vendor A (Best of Breed) and Vendor B was also minimal (0.5 rating difference) so Vendor B is again a solid choice.

BC/DR: Vendor B was the lowest rated solution for BC/DR, but again focusing on the customers requirements, the solution exceeded the minimum requirement of 3 comfortably with a rating of 4.2. Choosing Vendor B meets the requirements and likely avoids any interoperability and/or support issues, meaning a simpler overall solution.

Let’s think about some of the advantages for a customer choosing a standard platform for all workloads in the event a platform meets all requirements.

1. Lower Risk

Having a standard platform minimizes the chance of interoperability and support issues.

2. Eliminating Silos

As long as you can ensure performance meets requirements for all workloads (which can be difficult on centralized SAN/NAS deployments) then using a standard platform will likely lead to better utilization and higher return on investment (ROI).

3. Reduced complexity / Single Pane of Glass Management

Having one platform means not having to have SMEs in multiple technologies, or in larger organisations multiple SMEs per technology (for redundancy and/or workload) meaning reduced complexity, lower operational costs and possibly centralized management.

4. Lower CAPEX

This will largely depend on the vendor and quantity of infrastructure purchased, however many customers I have worked with have excellent pricing from a vendor as a result of standardizing.

Summary:

I am in no way saying “One size fits all” or that “every problem is a Nail” and recommending you buy a hammer. What I am saying is when considering infrastructure for your environment (or your customers), avoid tunnel vision and consider the other workloads or existing infrastructure in the environment.

In many cases the “Best of Breed” solution is not required and in fact implementing that solution may have significant implications in other areas of the environment.

In other cases, workloads may be so mission critical, that a best of breed solution may be the only way to meet the business requirements, in which case, a using a standard platform that may not meet the requirements would not be advised.

However if you can meet all the customer requirements with a standard platform while working within constraints such as budget, power, cooling, rack space and time to value, then I would suggest your doing yourself (or your customer) a dis-service by not considering using a standard platform for your workloads.

Related Articles:

1. Enterprise Architecture & Avoiding tunnel vision.

2. Peak Performance vs Real World Performance

Peak Performance vs Real World Performance

In this post I will be discussing Real World Performance of Storage solutions compared to peak performance. To make my point I will be using some car analogies which will hopefully assist in getting my point across.

Starting with the Bugatti Veyron Super Sport (below). This car has a W16 engine with 4 turbochargers and produces 1183BHP (~880kW) and has a top speed (peak performance) of 267MPH (431KPH).

bugatti-veyron-super-sport-

The Veyron achieved the world record 267MPH at Volkswagen’s Ehra-Lessien test track in Germany. The test track has a 5.6 mile long straight. This is one of the very few places on earth where the Veyron can actually achieve its peak performance.

Now for the Veyron to achieve the 267MPH, not only do you need a 5.6 mile long straight, but the Veyron’s rear spoiler must NOT be deployed. Now rear spoilers provide down-force to keep stability so having the spoiler down means the car has a reduced ability to for example take corners.

bugatti-veyron-super-sport_100315491_l

In addition to requiring a 5.6 mile long straight, the rear spoiler being down, the Veyron can also only maintain its top speed (Peak performance) for 12 minutes before the Veyron’s 26.4-gallon fuel tank will be emptied, which is lucky because the Veyron’s specially designed tyres only last 15mins at >250MHP.

veyron-tires-2-thumb-550x336

So in reality, while the Bugatti Veyron is one of (if not the fastest) production car in the world, even when you have all your ducks in a row, you can still only achieve its peak performance for a very short period of time (in this example <12 mins) and with several constraints such as reduced ability to corner (due to reduced aerodynamics from the spoiler being down).

Now what about Fuel Economy? The Veyron is rated as follows:

City Driving: 29 L/100 km; 9.6 mpg

Highway Driving: 17 L/100 km; 17 mpg

Top Speed: 78 L/100 km; 3.6 mpg

As you can see, vastly different figures depending on how the Veyron is being used.

There are numerous other factors which can limit the Veyron’s performance, such as weather. For example if the test track is wet, or has strong head winds, the Veyron would not be able to perform at its peak.

bugatti-veyron-wallpaper-7

So while the Veyron can achieve the 267MPH, In the real world, its average (or Real World) performance will be much lower and will vary significantly from owner to owner.

At this stage you’re probably asking “What has this got to do with Storage”?

A Storage solution, be it a SAN/NAS or Hyper-Converged, all can be configured and benchmarked to achieve really impressive Peak Performance (IOPS) much like the Veyron.

But these “Peak Performance” numbers can rarely (if at all) be achieved with “Real World” workloads, especially over an extended duration.

To quote two great guys in the Storage industry (Vaughn Stewart & Chad Sakac):

Absolute performance more often than not, is NOT the only design consideration.

I couldn’t agree with this more. The storage vendors are to blame by advertising unrealistic IOPS numbers based on 4K 100% read and now customers expect the same number of IOPS from SQL or Oracle.

The MPG of the Veyron is like the number of IOPS a Storage array can achieve. It Depends on how the Car or Storage Array is used! The car will get higher MPG if used only on the highway just like a Storage Array will get higher IOPS if only used for one I/O profile.

As the IO size and profile of workloads like SQL & Oracle are vastly different than the peak performance benchmarks using 4K 100% Read IOPS, expecting the same IOPS number for the benchmark and SQL/Oracle is as unrealistic as expecting the Veyron to do 267MPH in heavy traffic.

heavy-traffic-beirut-saidaonline

But like I said, Its the storage vendors fault for failing to educate customers on real world performance so many customers have the impression that peak IOPS is a good measurement, and as a result customers regularly waste time comparing Peak Performance of Vendor A and Vendor B, instead of focusing on their requirements and Real World performance.

In the real world, (at least in the vast majority of cases) customers don’t have dedicated storage solutions for one application where peak performance can be achieved, let alone sustained for any meaningful length of time.

Customers generally run numerous mixed workloads on their storage solutions, everything from Active Directory, DNS , DHCP etc which has low capacity/IOPS requirements , Database, Email and Application servers which may have higher capacity/IOPS requirements to achieve and backup with are low IOPS but high capacity.

Each of these workloads have different IO profiles and depending on storage architecture may share storage controllers / SSDs / HDDs / storage networking all of which can result in congestion / contention which leads to reduced performance.

Before you start considering what vendors storage solution is best, you need to first understand (and document) your requirements along with a success criteria which you can validate storage solutions against.

If your requirements are for example:

  • Host 10TB of Exchange Mailboxes for 2000 users (~400 random Read/Write 32-64k IOPS)
  • Host 20TB Windows DFS solution
  • Host 50TB of Backups
  • Support 1TB active working set SQL Database
  • Host 10TB of misc low IO random workload
  • Have Per VM snapshot / backup / replication capabilities

Then there is no point having (or testing) a solution for 100k Random Read 4k IOPS, as your requirement may be less than 10K IOPS of varying sizes and profile.

Consider this:

If the storage solution/s your considering can achieve the 10K IOPS with the I/O profile of your workloads and can be easily scaled, then a solution able to achieve 20K IOPS day 1, is of little/no advantage to a solution which can achieve 12K IOPS since 10K IOPS is all that you need.

Now if your Constraints are:

  • 12RU rack space
  • 4kw Power
  • $200k

Anything that’s larger than 12RU, uses more than 4Kw of Power or costs more than $200k is not something you should spend your time looking at / benchmarking etc since its not something you can purchase.

So to quote Vaughn and Chad again, “Don’t perform Absurd Testing”. absurdtesting

In my opinion, customers should value their own time enough not to waste time doing a proof of concepts (PoCs) on multiple different products when in reality only 2 meet your requirements.

An example of Absurd testing would be taking a Toyota Corolla on a test drive to a drag strip and testing its 1/4 mile performance when you plan to use the car to pick-up the shopping and drop the kids off at school.

school crossingcarshopping

Its equally as Absurd to test 100% Random Read 4k IOPS or consider/test/compare a storage solutions <insert your favourite feature here> when its not required or applicable to your use case.

Summary:

  1. Peak performance is rarely a significant factor for a storage solution.
  2. Understand and document you’re storage requirements / constraints before considering products.
  3. Create a viability/success criteria when considering storage which validates the solution meets you’re requirements within the constraints.
  4. Do not waste time performing absurd testing of “Peak performance” or “features” which are not required/applicable.
  5. Only conduct Proof of Concepts on solutions:
    1. Where no evidence exists on the solutions capability for your use case/s.
    2. Which fall within your constraints (Cost, Size , Power , Cooling etc).
    3. Which on paper meet/exceed your requirements!
    4. Where you have a documented PoC plan with a detailed success criteria!
  6. As long as the solution your considering can quickly, easily and non-disruptively scale, there is no need to oversize day 1.
    1. If the solution your considering CANT quickly, easily and non-disruptively scale, then its probably not worth considering.
  7. The performance of a storage solution can be impacted by many factors such as compute, network  and applications.
  8. When Benchmarking, do so with tests which simulate the workload/s you plan to run, not “hero” style 100% read 4k (to achieve peak IOPS numbers) or 100% read 256k (to achieve high throughput numbers).

How much detail to include in a VCDX Submission

Today I saw a conversation on Twitter where a VCDX candidate was asking about how much depth to go into for a VCDX submission (see tweet below)

vcdxdetail

I think this is a great question and one which I am asked regularly as a VCDX mentor. I responded with the follow tweets.

Tweet 1:

vcdxdetail3

Tweet 2:

vcdxdetail4

Since the information I provided was well received (see below), I thought a quick post was in order so others can benefit.

vcdxdetail2

My advise to VCDX candidates is as follows:

1. For any significant design decision made, document the decision in an Appendix in a similar format to the examples in my VMware Architectural Decision Register. This includes “Problem Statement”, “Assumptions” , “Motivation” , “Decision” , “Justification”, “Alternatives”, “Implications” and “Related Decisions”.

2. Include cross references in the design to the detailed design decision in the appendix.

3. Ensure each design decision has multiple “Alternatives”

4. Ensure you have thought through each “Alternative” and have a deep (and equal) understanding of the Pros/Cons of each.

5. Cross reference the design decision back to customer Requirement/s , Risk/s , Constraint/s , Assumption/s.

6. Ensure you know each of the design decisions like the back of your hand before the defence.

7. Pick 3 to 5 key design decisions to cover off in detail during the defence which allow you to demonstrate expertise, for example, a decision which goes against best practice and can be justified would be a good place to start.

Best of luck with your VCDX journey!

Related Articles:

1. My VCDX Journey

2. The VCDX Application process

3. The VCDX candidates advantage over the panellists.

4. VCDX Defence Essentials – Part 1 – Preparing for the Design Defence

5. VCDX Defence Essentials – Part 2 – Preparing for the Design Scenario

6. VCDX Defence Essentials – Part 3 – Preparing for the Troubleshooting Scenario